Skip to content
All work

Hardware/SoftwareUnited States2023

Crash observability and OTA for a Cortex-M fleet

Coredump capture, symbolication, and staged OTA that turn field failures into fixed builds.

Client
Memfault
Region
United States
Sector
Embedded observability + OTA
Engagement
2023 · 15 weeks
Team
2 firmware · 2 backend · 1 platform
Status
In production

The challenge

A connected-hardware maker was flying blind on field failures — crashes surfaced only as returns, with no root cause and no safe way to push a fix. They needed on-device fault capture plus a backend for symbolication, metrics, and staged OTA.

What we built

The full pipeline, end to end — not a blurb.

  1. 01Firmware

    A Cortex-M fault handler capturing coredumps and metrics with compact upload and watchdog integration.

  2. 02Backend

    A symbolication service, issue de-duplication, device metrics, and OTA release orchestration.

  3. 03Product

    A crash explorer and fleet-health view with staged rollout and automatic halt.

  4. 04Operations

    Kubernetes, object storage, and alerting.

Results

20 min

crash root-cause time via symbolicated traces, from 3 days

4.2 KB

median coredump upload overhead per event

0

bricked units across staged OTA rollouts

61%

lower field crash rate over three release cycles

Now the release gate for every firmware ship.

Have a constraint like Memfault’s? Bring us yours.

Start a project

Next engagement

PriceOye

E-commerce web + operations